This paper presents a probabilistic methodology in evaluating corrosion defect characteristics of carbon steels offshore pipelines. A nonlinear multivariate regression model was selected to describe the correlation among corrosion defect parameters while the least-squares method was used to minimize its residuals. The proposed framework were able to provide better insights on the degree of correlations among corrosion defect parameters, which eventually proven that the interactions among defects are indeed significant. © (2013) Trans Tech Publications, Switzerland.